CN101615934B - Full diversity high-rate coded transmit beamforming techniques using partial-algebraic precoding, as well as associated near-optimal low-complexity receivers - Google Patents

Abstract

The present invention relates to full diversity high-rate coded transmit beamforming techniques using partial-algebraic precoding, as well as associated near-optimal low-complexity receivers. The present invention concerns a system for transmitting data from a transmitter comprising at least two transmitting antennas, an encoder (ENC) following an error correcting code structure and a transmit beamforming technique to a receiver comprising at least two receiving antennas and a decoder defined according to the error correcting code structure over a transmit beamformed MIMO channel, characterised in that the coding rate Rc of the encoder (ENC), the parameters of said transmit beamformed MIMO channel and a target diversity order of the system are selected dependently one from each other. The invention also relates to an apparatus for transmitting data to a receiver over a transmit beamformed MIMO channel.

Description

High-rate coded launching beam forms technology and near optimal low complex degree receiver

Technical field

The present invention relates generally to that the mimo channel (transmit beamformed MIMO channel) for being formed by launching beam transmit data to the system of receiver from transmitter, described transmitter comprises at least two transmitting antennas, the encoder of follow error correcting code structure (error correcting code structure) and launching beam formation technology, the decoder that described receiver comprises at least two reception antennas and limits according to described error correcting code structure.

Background technology

Hereinafter, channel resource some Spatial Dimensions (spatial dimension) of meaning frequency bandwidth, the time interval and being brought by the transmitting antenna and reception antenna that are placed on different locus possibly.

Due to channel variation (such as decay (fading) and cover (shadowing)), the signal launched by wireless channel suffers serious deterioration, and this permission is used as channel as stochastic variable.Hereinafter, the time required for launching about information word considers slow channel variation, but the realization of channel should change between the twice emitting of information word.Preventing from being called as a kind of important method that quasistatic (quasi-static) decays is be provided in the diversity (diversity) in time, frequency or space.

Channel diversity order is restricted to the quantity of the independence decay stochastic variable observed in the channel resource for transmitting.Transmission/reception scheme can collect the specified rate of diversity, is called as the order of diversity (diversity order) of system, determines its upper bound, be also called full-diversity degree by channel diversity order.

When information word is received the incorrect estimation of machine, just there occurs error event.The principal parameter calculating the probability of mistake be associated with given error event is allowed to be Euclidean distance (Euclidean distance) between the noiseless received signal be associated with the information word of launching and the noiseless received signal of reconstruction be associated with the information word of decoding.The order of diversity of error event is restricted to the quantity of the independent random variable related in the Euclidean distance be associated with error event.Finally, system diversity degree equals the minimum order of diversity of all possible error event, or is equivalent to all possible information word pair.

The telecommunication system of multiple antenna is used to be called multi-input multi-output system (being also referred to as mimo system) at the receiver end of wireless link and/or transmitter terminal.Mimo system is shown as relative to the transmission capacity that single antenna system provides and provides larger transmission capacity (transmission capacity).Particularly, for given signal to noise ratio with under suitable uncorrelated channel condition, MIMO capacity along with launch or reception antenna quantity and increase, be whichsoever minimum.MIMO technology thus probably for Future wireless systems, this wireless system is intended to provide large spectrum efficiency or alternately reduces and obtains transmitting power spectrum efficiency (spectral efficiency) required for suitable with the spectrum efficiency that obtains in current telecommunication systems.MIMO technology probably combines with the multi-carrier modulation technology being similar to OFDM, and this allows mimo channel model to be thought of as do not have intersymbol interference.

The transmitter of mimo system comprises digital modulator, this digital modulator is used for spatial flow coded bit stream (coded bits stream) being converted to modulation symbol, the input of this digital modulator is coded-bit, and the output of this digital modulator is N _s≤ min (N _t, N _r) vector of individual modulation symbol.This system launches N subsequently on transmitting mimo channel _sindividual spatial flow.

Characteristic vector launching beam forms the performance that scheme (eigenvector transmit beamforming scheme) can be used for improving mimo channel.The matrix of the mimo channel that launching beam is formed is the N of this channel _t× N _rmatrix H and N _s× N _tthe cascade (concatenation) of beam forming matrix.Described N _s× N _tbeam forming matrix is from the N with channel _t× N _rthe N of matrix _sthe characteristic vector that individual best features value (best eigenvalue) is associated draws.Subsequently, launching beam forms technology needs the incomplete knowledge of the matrix H at transmitter place to apply pre-filtering (pre-filtering), and this pre-filtering allows the low complex degree optimum monitoring at receiver place.Such as, incomplete knowledge be manage at receiver everywhere and the quantised versions of the channel estimating fed back from receiver to transmitter.

Assuming that during the spatial flow of single transmitting, characteristic vector launching beam forms scheme and reaches full-diversity degree.When launching multiple spatial flow when the data rate in order to increase the program, the order of diversity of the program significantly reduces.

The object of the invention is to obtain high data rate while maintenance receiver low complex degree, and ensure target diversity and high performance.In fact, at transmitting pusher side, diversity can be overcome by the algebraic linear precoder (algebraic linear precoder) comprising linear combination (combine) modulation symbol and reduce.But at receiver side, such algebraic linear precoder generally includes high complexity and detects.

In addition, inventor notices that comprising FEC encoder at transmitter place allows to benefit from the partial diversity that multiple transmitting antenna brings, and this depends in part on the code rate (coding rate) of FEC encoder.Inventor is also noted that application algebraic linear precoder contributes to recovering (recover) order of diversity part of being brought by multiple transmitting antenna, this depends in part on precoding size (precoding size), is restricted to the quantity of combined modulation symbol.By at transmitter place composition error correcting code and linear precoder, can independently select the code rate of error correction and precoding size to observe target diversity order at receiver place, such as full-diversity degree.

The object of the invention is to determine the relation between the parameter of the mimo channel that precoding size, launching beam are formed and accessible order of diversity.Then, select precoding size for obtaining the target diversity order of system.

Because the complexity of detector increases along with precoding size, the such relation between the parameter of mimo channel, error correction bit rate and target diversity order formed at precoding size, launching beam contributes to finding minimum precoding size, and described minimum precoding size allows the Receiver Complexity of realize target order of diversity and reduction.

In addition, another object of the present invention utilizes algebraic simplification (algebraic reduction) technology to reduce the complexity of detector when keeping identical point of set attribute.Under this assumption, if precoding size is minimized, then the performance of detector is optimised.

Summary of the invention

The present invention relates to the system for being transmit data to receiver from transmitter by mimo channel, described transmitter comprises at least two transmitting antennas, follows the encoder (ENC) of error correcting code structure and launching beam formation technology, the decoder that described receiver comprises at least two reception antennas and limits according to described error correcting code structure, the system is characterized in that the selection of the parameter of mimo channel that the code rate of encoder, described launching beam are formed and the target diversity order of system is complementary.

Described receiver comprises detector, and the output of described detector is the estimation to coded-bit.

By assemble described digital modulator, described launching beam formed mimo channel and described detector limit equivalent channel model.The input of described equivalent channel model is coded-bit, and the output of described equivalent channel model is to the soft estimation of coded-bit or firmly estimates.The channel model of described equivalence be decomposed at receiver place (decompose) comprise nested block attenuation channel.

The Mathematical Modeling of nested block attenuation channel is made up of the cascade of the block of equivalent channel coefficient (coefficient).A coded-bit is assumed on a block to be launched.All coded-bits that same block is launched experience identical equivalent channel coefficient.Each equivalent channel coefficient is the combination of the one or several realizations of the stochastic variable being with given order of diversity.Equivalent channel has five parameters, is respectively: the quantity N of spatial flow _s, the quantity N of transmitting antenna _t, the quantity N of reception antenna _r, the set D={N of the order of diversity that the equivalent channel coefficient with each piece is associated _tn _r, (N _t-1) (N _r-1) ..., (N _t-N _s+ s) (N _r-N _s+ s) }, and the lengths sets LB={LB of block (1) ..., LB (N _s), the quantity of cascade block equals N _sand wherein L is the quantity of the coded-bit of each code word, and each piece is associated with the spatial flow i launched.The length LB (i) of each piece is derived from the quantity of the bit of each modulation symbol be associated with the i-th spatial flow.The decay stochastic variable be associated with i-th piece is by subset ∑ (i) of independent random variable D (i) and integer value D (i)≤N _tn _rcombination limit, make its implication is d (i)>=D (j), and supposition D (1)=N _tn _rthere is the highest order of diversity.The code rate R of described encoder _c, five parameters of described nested block attenuation channel and accessible aims of systems order of diversity δ be by following relational links:

δ=D (i), wherein i by $\underset{k=1}{\overset{i-1}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)<R_c\underset{k=1}{\overset{\left|D\right|}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)\&le;\underset{k=1}{\overset{i}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)$ Provide,

Wherein | D| is the radix of set D.

According to the feature of transmitter, transmitter comprises algebraic linear precoder further, and the input of described algebraic linear precoder is N _sthe vector of individual modulation symbol, it exports the input being given to the mimo channel that launching beam is formed.Described algebraic linear precoder has parameter s, and this parameter s is called as precoding size, and it meets s≤N _s, in order to by N _ss modulation symbol linear combination of the identical vector of individual modulation symbol is together to produce the modulation symbol of s combination.It provides the modulation symbol and N that comprise s combination _sthe N of-s the modulation symbol do not combined _sthe output vector of individual pre-coded symbols, is associated with a transmission of the mimo channel formed by launching beam.

Described linear precoder is by N _sthe input vector of individual modulation symbol is multiplied by N _s× N _smatrix represents.

According to the execution mode of algebraic linear precoder, the pre-coded symbols of described output vector is obtained by the vector of modulation symbol and the product of complex linear matrix S, and described complex linear matrix S is provided by following:

$S=P_1\left[\begin{array}{cc}{S}^{\&prime;}& 0\\ 0& I\end{array}\right]P_2$

Wherein P ₁and P ₂n _s× N _spermutation matrix, and S ' meets following s × s matrix, if vector and the matrix S ' be multiplied of i.e. s modulation symbol, sent by the diagonal band noise Rayleigh fading channel being of a size of s × s, and decoded by maximum likelihood decoder, then the order of diversity of performance equals s.

Preferably, permutation matrix P ₁be selected as equaling unit matrix, permutation matrix P ₂be chosen as symbol [X (1), the X (N of the combination making output vector _s-s+2) ..., X (N _s)] be the modulation symbol [Z (1) of the described vector of modulation symbol, ..., Z (s)] linear combination, other pre-coded symbols of such output vector is the symbol that do not combine and meets [X (2), ..., X (N _s-s+1)]=[Z (s+1) ..., Z (N _s)].

Under this selection of linear pre-coding scheme, equivalent nested channel has five parameters, respectively: the quantity N of spatial flow _s≤ min (N _t, N _r), the quantity N of transmitting antenna _t, the quantity N of reception antenna _r, the set D={N of the order of diversity that the equivalent channel coefficient with each piece is associated _tn _r, (N _t-1) (N _r-1) ..., (N _t-N _s+ s) (N _r-N _s+ s) }, and the lengths sets of block $\mathrm{LB}=\{\mathrm{LB}\left(1\right)+\underset{i=N_s-s+2}{\overset{N_s}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(i\right),\mathrm{LB}\left(2\right),...,\mathrm{LB}(N_s-s+1)\},$ The quantity of cascade block equals N _s-s+1 and wherein L is the quantity of the coded-bit of each code word.

The code rate R of encoder _c, five parameters of described nested block attenuation channel and accessible aims of systems order of diversity δ be by following relational links:

δ=D (i), wherein i by $\underset{k=1}{\overset{i-1}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)<R_c\underset{k=1}{\overset{\left|D\right|}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)\&le;\underset{k=1}{\overset{i}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)$ Provide,

Note as s=1, mean and do not have modulation symbol to be grouped together, then algebraic linear precoder is included in the nested channel of equivalence of system is wherein that transmitter does not comprise the nested channel of same equivalence of the system of algebraic linear precoder wherein.

According to the characteristic of system, identical modulation is used for each spatial flow, that is, lB (i)=LB (j)=L/N _s.The diversity δ (s) observed in the output of receiver is by providing as follows:

Receiver comprises detector, the output of described detector is the estimation to coded-bit, described detector comprises the first detector block and the second detector block, described first detector block is intended to the coded-bit estimating to be associated with s the modulation symbol combined of the vector of the modulation symbol carried by the vector received, and described second detector block is intended to estimate the N with the described vector of modulation symbol _sthe coded-bit that-s the modulation symbol do not combined is associated.

Preferably, described first detector block allows to recover the order of diversity that brought by described algebraic linear precoder S ', and described second detector block is to provide the linear detector of soft output estimation to coded-bit or hard output estimation.

According to the characteristic being intended to the described receiver providing low complex degree to detect, the vector received at receiver place is provided by following:

Wherein Δ is rectangle diagonal angle N _t× N _rmatrix, its diagonal values is the relevant non-identity distribution stochastic variable sorted with size descending; And N ₂it is additional white Gauss noise vector.The vector of described reception is subsequently by given as follows:

Wherein Z=[Z '; Z "] is the vector of modulation symbol, and D is s × s diagonal matrix, and D ' is (N _s-s) × (N _r-s) diagonal angle rectangular matrix, and N ' and N " is noise vector.Subsequently, the first detector block converts Z ' S ' D+N ' to estimation to the coded-bit be associated with Z ', and described second detector block Z " D '+N " is converted to the Z " estimation of the coded-bit be associated.

As described below, by using algebraic linear to simplify the complexity reducing the first detector block.

The mimo channel that the invention still further relates to for being formed by launching beam transmit data to the device of receiver, and described receiver comprises at least two and receives sky.Described device comprises at least two transmitting antennas, follows the encoder of error correcting code structure and digital modulator, and the output of described digital modulator is N _sthe vector of individual modulation symbol.Described receiver comprises the decoder limited according to error correcting code structure, and this device characteristic is that it also comprises algebraic linear precoder, and the parameter s of this algebraic linear precoder is called as precoding size, and it meets (verify) s≤N _s.This algebraically precoder is intended to by s modulation symbol linear combination of the identical vector of modulation symbol together, be called hybrid modulation symbol, and is intended to the output vector providing the pre-coded symbols be associated with a transmission of the mimo channel formed by launching beam.

The channel that the invention still further relates to for being formed by launching beam receives the device of data vector from the transmitter comprising at least two transmitting antennas.Described device comprises at least two reception antennas, according to the decoder that error correcting code structure limits, output is the detector of the estimation to coded-bit, some components (component) in the vector of described reception are the linear combination of some modulation symbols in the identical vector of armed modulation symbol, this device characteristic is: described detector comprises the first detector block and the second detector block, described first detector block is intended to estimate and the coded-bit that the modulation symbol of the vector of the modulation symbol carried by the described data vector received involved in described linear combination is associated, described second detector block is intended to the coded-bit estimating to be associated with the modulation symbol of the described vector of the modulation symbol do not related in described linear combination.

The linear combination of modulation symbol is obtained by algebraic linear precoder at transmitting pusher side, this device characteristic is: described first detector block allows the order of diversity recovering to be brought by described algebraic linear precoder, and described second detector block is to provide the linear detector of soft output estimation or hard output estimation.

Accompanying drawing explanation

By the reading of the explanation to following instance execution mode, characteristic of the present invention will clearly, described explanation with reference to appended accompanying drawing, wherein:

Fig. 1 represents the scheme of the data transmission system of the mimo channel by Wave beam forming;

Fig. 2 represents the scheme of the equivalent channel that binary system input binary system exports;

Fig. 3 represents the scheme of the block cascade of nested block attenuation channel;

Embodiment

Data transmission system SYST comprises transmitter TRD and receiver RCV, and described transmitter TRD comprises N _t>=2 transmitting antennas, described receiver RCV comprises Nr>=2 reception antenna and channel.

Described transmitter TRD comprises the encoder ENC, the digital modulator MOD that follow error correcting code structure, the beam-forming device TxBF implementing launching beam formation technology and comprises algebraic linear precoder ALP according to preferred embodiment.

The decoder DEC that this receiver RCV comprises detector DET and limits according to this error correcting code structure.

Described transmitter TRD can be such as base station, and receiver RCV is such as mobile subscriber equipment.

Saying of summary, carries out Digital Transmission process: { b} is with speed R for armed information data bits as follows _c=K/L feeds encoder ENC, and wherein K is the input data bit { quantity of b}, and L is output codons { c} amount of bits.Described encoder ENC can follow the error correcting code structure of any type, such as, LDPC (low-density checksum (Low Density Parity Check)) code, special rich code (turbo-code), block code (block-code) (such as Reed-Solomon), Binary Convolutional Code (binary convolutional code) etc.

The modification of transmission system utilizes bit interleaver (bit interleaver) INT that coded-bit is interlocked, and utilize bit deinterlacer (bit de-interleaver) DINT to apply the bit deinterleave be associated at receiver side, as shown in Figure 1.According to error correcting code structure design bit interleaver INT to guarantee to show obtainable target diversity order in the output performance of the decoder DEC be associated.

Described digital modulator MOD can be BPSK (binary phase shift keying) or QPSK (orthogonal PSK) modulation, be such as preferably 2 " '-quadrature amplitude modulation (2 " '-QAM).The input of described digital modulation MOD is that { { c}, its output is modulation symbol to coded-bit, and this modulation symbol defines N for c} or the staggered version according to this coded-bit of system variation _sthe vector Z of individual modulation symbol Z (i).For N _seach modulation of applying in individual symbol is not necessarily identical, and can have different input bit quantity.

Described algebraic linear precoder ALP is mathematically by N _s× N _scomplex matrix (complex matrix) S represent.

N _svector Z and the N meeting following character of individual modulation symbol _s× N _tmatrix S is multiplied:

$S=P_1\left[\begin{array}{cc}{S}^{\&prime;}& 0\\ 0& I\end{array}\right]P_2$

Wherein P ₁and P ₂it is displacement (permutation) matrix, S ' is s × s matrix, it meets following character: if the vector of s modulation symbol is by matrix S ' precoding, sent by the diagonal band noise Rayleigh fading channel (diagonal noisy Rayleigh fading channel) being of a size of s × s, and decoded by maximum likelihood decoder, then the order of diversity of performance equals the size s of described square matrix S ', is called precoding size.In other terms, S ' is called as full-diversity linear precoder.

The output of described precoder ALP is N _sthe output vector X of individual pre-coded symbols, s wherein are the linear combination of modulation symbol Z (i) of vector Z, are called the modulation symbol of combination, and other N _sindividual modulation symbol Z (i) equaling same vector Z of-s, is called the modulation symbol do not combined.

Symbolic vector X is given as the input that described launching beam forms scheme subsequently, and it exports N _tthe vectorial X ' of individual symbol, the input of described detector DET is N _rvectorial Y, the Y of individual symbol are provided by following:

Y＝X′H+N

Wherein N is length N _radditive White Gaussian Noise vector (additive white Gaussian noise vector), H is the N of channel _t× N _rmatrix, its yuan (entry) is independently answer Gaussian random variable (independent complex Gaussian random variable).

Apply unusual (singular) value to H to decompose, and select its N _sindividual best features value, the channel matrix H of channel is written as:

H＝UΔ _HV (1)

Wherein U is N _t× N _tunitary matrice (unitary matrix)

V is N _r× N _runitary matrice

And Δ _hdiagonal angle rectangle N _t× N _rmatrix, its diagonal values is the min (N of (sorted in decreasing order of magnitude) of sorting with size descending _t, N _r) individual singular value.This singular value is relevant non-identity distribution stochastic variable (dependent non-identically distributed random variable).

Suppose that the H by equation (1) is given is known, described by previous section at transmitter.N _tthe vectorial X ' of individual pre-coded symbols can be written as X '=XT, and wherein beam forming matrix T is by U ⁺n _sthe N that individual the first row builds _s× N _tmatrix, and U ⁺it is the transpose conjugate (transpose conjugate) of matrix U.That is, T is at the N with H _sprojection (projection) matrix on the subspace that individual best singular value is associated.The vector received is write again:

Wherein Δ is the N of H _sthe N of individual best singular value _s× N _rdiagonal matrix, and N ₂it is Additive White Gaussian Noise vector.

I-th singular value has the order of diversity (N be associated _t-i+1) (N _r-i+1).Further, any two singular values are to having non-zero correlation.

As described above with respect to Fig. 2, digital modulator MOD, algebraic linear precoder ALP, beam-forming device TxBF, by transmitting and receiving the channel of antenna and the gathering of detector DET define equivalent channel model, described equivalent channel model is called as binary system input binary system output equivalent channel BIBOCH, the input of described channel is coded-bit, and it exports and is to the soft estimation of coded-bit or firmly estimates.

Due to error-correcting code ENC structure with alternatively due to interleaver, the coded-bit seen at receiver place is just as these coded-bits are launched by binary system input binary system output equivalent channel BIBOCH, described binary system input binary system output equivalent channel BIBOCH can be believed to comprise binary modulated device BM and nested block attenuation channel NBFCH, just as shown in Figure 1.Described binary modulated device BM applies (scaled) BPSK modulator of scale, that is, the value (opposite value) (such as, A and-A) that ' 0 ' bit value is contrary with two respectively with ' 1 ' bit value is associated.Further, corresponding Euclidean distance 2A can change to the transmission of another bit according to a bit, and it mainly relies on digital modulation MOD binary map.

When not having a linear predictive coding (s=1), the Mathematical Modeling forming the nested block attenuation channel NBFCH of (as shown in Figure 3) with the cascade of block has five parameters, is respectively: the quantity N of spatial flow _s≤ min (N _t, N _r), the quantity N of transmitting antenna _t, reception antenna quantity N _r, the order of diversity set D={N to be associated with the equivalent channel coefficient of each piece _tn _r, (N _t-1) (N _r-1) ..., (N _t-N _s+ 1) (N _r-N _s+ 1) } and the lengths sets LB={LB of block (1) ..., LB (N _s).The quantity of cascade block equals N _sand wherein L is the quantity of the coded-bit of each code word.The length LB (i) of each piece is derived from the quantity of the bit of each modulation symbol be associated with the i-th spatial flow, and the decay stochastic variable be associated with i-th piece is by subset ∑ (i) of independent random variable D (i) and integer value D (i)≤N _tn _rcombination limit, make (its implication is d (i)>=D (j)), and supposition D (1)=N _tn _rthere is the highest order of diversity.

Follow such equivalent channel model, described decoder DEC is by coded-bit { the reception version of c} be converted to information data bits { the soft estimation of b}

According to the characteristic of transmitter TRD, the code rate R of encoder ENC _c, five parameters of described nested block attenuation channel NBFCH and obtainable aims of systems order of diversity δ, there is following relation:

δ=D (i), wherein i by $\underset{k=1}{\overset{i-1}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)<R_c\underset{k=1}{\overset{\left|D\right|}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)\&le;\underset{k=1}{\overset{i}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)$ Provide,

Wherein | D| is the radix (cardinality) of set D.

Equation above establishes the code rate R of target diversity order in singular value D (i) quantity relative to channel, encoder ENC _c, channel five parameter (quantity N of spatial flow _s, transmitting antenna quantity N _t, reception antenna quantity N _r, to be associated with the equivalent channel coefficient of each piece order of diversity set D, block lengths sets LB) between relation.

Note, in practice, modulation symbol Z (i) may by except other the channel imperfections except the channel imperfections that is associated of decay stochastic variable disturb, such as, the noise added, other the stochastic variable of multiplication.Important explanation is, add that (plus) brings other stochastic variables of additional order of diversity compared to that being associated with the transmission of Z (i-1), it is identical stochastic variable collection that generation collects with the stochastic variable (singular value) of the order of diversity that Z (i) is associated.

According to the embodiment of algebraic linear precoder ALP, permutation matrix P ₁selected equal unit matrix (identity matrix), selective cementation matrix P ₂make pre-coded symbols [X (1), the X (N of output vector _s-s+2) ..., X (N _s)] be the hybrid modulation symbol [Z (1) of described modulation symbol vector, ..., Z (s)] linear combination, other pre-coded symbols of such output vector is the symbol that do not combine and meets [X (2), ..., X (N _s-s+1)]=[Z (s+1) ..., Z (N _s)].This embodiment improves the order of diversity observed at decoder output.

Inventor has noticed that the channel that launching beam is formed is orthogonal, this means when without any linear predictive coding, reach by performing each dimension (per-dimension) of managing at detector DET everywhere the optimum detection of the signal received, described detector DET generation is to the soft estimation of coded-bit or firmly estimate.

The linear precoding of present hypothesis (i.e. s > 1), best detection is the modulation symbol [Z (1) in combination, ..., Z (s)] united to reach when detecting, if such as use soft output, by using exhaustive marginalisation (exhaustive marginalization), these describe below.Such characteristic of described detector allows the diversity obtaining the performance observed in the output of receiver, and it is provided by following subsequently:

δ (s)=D (i), wherein i by $\underset{k=1}{\overset{i-1}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)<R_c\underset{k=1}{\overset{\left|D\right|}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)\&le;\underset{k=1}{\overset{i}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(k\right)$ Provide

Wherein D={N _tn _r, (N _t-1) (N _r-1) ..., (N _t-N _s+ s) (N _r-N _s+ s) } and

$\mathrm{LB}=\{\mathrm{LB}\left(1\right)+\underset{i=N_s-s+2}{\overset{N_s}{\mathrm{\&Sigma;}}}\mathrm{LB}\left(i\right),\mathrm{LB}\left(2\right),...,\mathrm{LB}(N_s-s+1)\}.$

Equation above establishes the code rate R of target diversity order in singular value D (i) quantity relative to channel, encoder ENC subsequently _c, channel five parameter (quantity N of spatial flow _s, transmitting antenna quantity N _t, reception antenna quantity N _r, to be associated with the equivalent channel coefficient of each piece order of diversity set D, block lengths sets LB), relation between precoding size s.

Equation above establishes the code rate R of described encoder ENC subsequently _c, spatial flow quantity N _s, transmitting antenna quantity N _t, reception antenna quantity N _rbe phase cross-correlation with precoding size s, and can be selected to obtain the target diversity order of system.

The growth of precoding size s adds the diversity perceived by decoder compared with the common beam-forming technology without linear predictive coding (s=1), this proves one of advantage of described system.

According to the characteristic of system, identical modulation is used for each spatial flow, that is: lB (i)=LB (j)=L/N _s.The diversity δ (s) observed in the output of receiver is by providing as follows:

Preferably, described precoding size s is selected to obtain target diversity order δ according to following equation _t:

$s=\underset{s^{\&prime;}}{\arg \min }\{\mathrm{\&delta;}\left(s^{\&prime;}\right)={\mathrm{\&delta;}}_t\}$

Described detector DET provides the soft estimation of the coded-bit be associated with the vector Z of modulation symbol Z (i) carried by the vectorial Y received or firmly estimates, as mentioned above.

When detector DET provides the hard output of coded-bit, optimum detection performance obtains from maximum likelihood (ML) detector, this is when the exhaustive search carried out on all possible candidate vector Z when maximizing likelihood probability (1ikelihood probability) p (Y|Z), namely to find symbol according to following

$\tilde{Z}=\arg \min \underset{z}{p}\left(Y\right|Z)$

Or to find the symbol minimizing quality factor (figure of merit), it is such as Euclidean distance || Y-ZSTH|| ²

$\tilde{Z}=\underset{z}{\arg \min }\left({\left|\right|Y-\mathrm{ZSTH}\left|\right|}^2\right)$

Subsequently from symbol obtain the estimation to the coded-bit be associated with s composite symbol

When described detector DET provides the soft output to coded-bit, when the exhaustive retrieval undertaken on all possible candidate vector Z by maximization figure of merit function (figure of merit function), (described maximization figure of merit function is each coded-bit C be associated with the vector Z of modulation symbol Z (i) _jminimise false probability) time, obtain optimum detection performance from MAP detector.Such as, described estimation is external probability, or A posterior probability (A Posteriori Probability).Maximum A posterior probability (MAP) detector allows to recover order of diversity δ (s).Relative to coded-bit C _jposterior probability (APP) utilize exhaustive marginalisation below to calculate:

$\mathrm{APP}\left(c_j\right)=\frac{\underset{Z\&Element;\mathrm{\&Gamma;}(c_j=1)}{\mathrm{\&Sigma;}}P\left(Y\right|Z)\underset{j}{\mathrm{\&Pi;}}\mathrm{\&pi;}(c_k=f(k,Z))}{\underset{Z\&Element;\mathrm{\&Gamma;}}{\mathrm{\&Sigma;}}P\left(Y\right|Z)\underset{j}{\mathrm{\&Pi;}}\mathrm{\&pi;}(c_k=f(k,Z))}---\left(2\right)$

Wherein, wherein N ₀noise variance (noise variance), Γ (c _j=1) be possible candidate vector Z collection, its mark has the i-th bit equaling 1, and Γ is the collection of all possible candidate vector Z, and f (k, Z) is the value of kth bit in the binary marks be associated with potential candidate vector Z.To the estimation of the coded-bit be associated with the vector Z of modulation symbol Z (i) carried by the vectorial Y received subsequently from relative to each coded-bit C _jmaximum A posterior probability (APP) in obtain.Note, relative to coded-bit C _jsoft output estimation be also often represented as log-likelihood ratio (LLR), it is by providing as follows:

${\mathrm{LLR}}_j=\log \left(\frac{\mathrm{APP}(c_j=1)}{\mathrm{APP}(c_j=0)}\right)$

Marginalisation be individual possible candidate vector Z carries out, and along with mN _srapid development become be difficult to process (intractable).

According to the modification of detector DET, to the described soft output estimation of the coded-bit be associated with s composite symbol consider be associated with the coded-bit of described s hybrid modulation symbol and the prior probability π (c provided by the output of decoder (DEC) _j).

According to another characteristic of receiver RCV, described detector DET is the detector of near optimal (near-optimal) of the low complex degree of the vectorial Y received, its produce to the N carried by the vectorial Y received _sthe estimation of the coded-bit that the vector Z of individual modulation symbol Z (i) is associated.If this detector can allow to obtain the performance showing the order of diversity identical with optimal detector after the decoding, then described detector is near optimal.

According to the execution mode of the detector DET of such near optimal, described detector DET comprises the first detector block DET1 and the second detector block DET2, described first detector block is intended to the coded-bit estimating to be associated with s the modulation symbol combined of the vector Z of modulation symbol Z (i) carried by vectorial Y, and described second detector block is intended to estimate the N with the vector Z of modulation symbol _sthe coded-bit that-s the modulation symbol do not combined is associated.Note, any symbol do not combined all independent of any other combination or the symbol that do not combine detect.

Further, the output estimation of the described detector block DET1 output estimation independent of detector block DET2 is obtained, and without any performance degradation (degradation).

Further, described detector block DET2 process N _sthe N of-s the symbol do not combined _s-s independent detection.But, in order to the near optimal of the optimality and more specifically detector DET that keep detector block DET1, still need to carry out joint-detection to s composite symbol.

According to the characteristic of the first detector block DET1, described first detector block DET1 allows the order of diversity recovering to be brought by described algebraic linear precoder ALP (namely by matrix S '), for the coded-bit with described s the estimation that hybrid modulation symbol is associated.Note, when matrix is S ' unit matrix (namely not having linear predictive coding), described first detector block DET1 provides the estimation to the coded-bit that first modulation symbol Z (1) of the vector Z with modulation symbol is associated.

According to the execution mode of the first such detector block DET1, to the estimation of described s the coded-bit that hybrid modulation symbol is associated preferably based on by following imperfect (incomplete) likelihood probability p (Y provided _{s, 0}| Z _{s, 0}) the calculating of function:

Wherein N ₀noise variance, Z _{s, 0}be the vector Z of modulation symbol, wherein said non-composite symbol equals 0, i.e. Z _{s, 0}=[Z (1) ..., Z (s), 0 ..., 0], and Y _{s, 0}=[Y (1) ..., Y (s), 0 ..., 0] be the vector received, wherein last (N _r-s) individual value equals 0.

Then, by selecting the minimum s≤N allowing to obtain target diversity _svalue, the complexity of such near optimal detector DET is minimized.In fact, only 2 of list are such as belonged to ^msindividual point for calculate soft output instead of

According to the execution mode of described second detector block DET2, described second detector block DET2 is the linear detector that decision function (decision function) Dec (.) follows.Described linear detector (preferably for hard output estimation) utilizes ZF (Zero Forcing (zero forces)) linear equalizer to provide soft output estimation or hard output estimation.Linear detector is to the symbol utilizing linear transformation (i.e. matrix multiplication) received.When hard output valve, decision function Dec (.) allows the receiving symbol of estimation to be converted to bit sequence.

As described earlier, the advantage of system increases precoding size s≤N _sadd the diversity of described decoder perception unfortunately, when described first detector block DET1 allows the order of diversity recovering to be brought by described algebraic linear precoder ALP, the complexity of described detector DET increases along with the value of assembled dimension s.Contrary, simple linear detector is proved to be and can not utilizes by the potential order of diversity brought of S.This defect overcome owing to employing algebraic linear precoder, its allow receiver place perform " algebraic lattice simplification " technology.Such receiver allow for the detector of low complex degree and ensure that the character of order of diversity.

According to the execution mode of such low complex degree detector, detector DET comprises following " algebraic lattice simplification " technology limited:

Make Δ '=P ₁Δ, and

${\mathrm{\&Delta;}}^{\&prime;}=\left[\begin{array}{cc}D& 0\\ 0& D^{\&prime;}\end{array}\right]\&DoubleLeftRightArrow;\mathrm{S\&Delta;}=\left[\begin{array}{cc}{S}^{\&prime;}D& 0\\ 0& D^{\&prime;}\end{array}\right]$

The vectorial Y of described reception then has following expression:

Wherein Z=[Z '; The vector that Z "] is modulation symbol, and D and D ' is diagonal angle rectangular matrix, and N ' and N " is noise vector.

According to this execution mode, described first detector block DET1 calculates the Z ' S ' D+N ' conversion to the estimation to the coded-bit be associated with Z ', and described second detector block DET2 calculating Z " D '+N " to the Z " conversion of the estimation of the coded-bit be associated.

Suppose, described s × s matrix D is decomposed in the following way:

D＝ΩΨΦ

Wherein Ω is s × s diagonal matrix, makes S ' Ω=T _us ', and T _us × s substrate variations (basis change) matrix (this substrate variations entry of a matrix is complex integers), diagonal angle s × s matrix Ψ has positive real diagonal element, and diagonal angle s × s matrix Φ has the complex element of unimodular (unity modulus).Such as, such decomposition is possible, condition is S ' is that cyclotomy rotates (cyclotomic rotation), described cyclotomy rotation description is in G.Rekaya, " the A Very Efficient Reduction Tool on Fast Fading Channels " of J-C.Belfiore and E.Viterbo, IEEE International Symposium on Information Theory and its Applications (ISITA), Parma, Italy, October 2004.

At receiver place, Ω is selected with optimization system performance, and matrix F is applied to the signal YV of reception ⁺, wherein:

At the vector of the reception of institute's filtering expression formula in, described first detector block DET1 compute vector to the hard estimation of the coded-bit be associated with Z ' or the conversion of soft estimation, as mentioned above, and described second detector block DET2 calculate Z " D '+N " to Z " the hard estimation of the coded-bit be associated or the conversion of soft estimation, as mentioned above.

Therefore, " algebraic lattice simplification " technology is on the second detector block DET2 without any impact, and it can select any type, such as linear soft output detector or hard output detector minimizing the complexity of described second detector block DET2.

For described first detector block DET1, inventor observes wherein be the collection of the complex vector of length s, its yuan has integer real part and imaginary part.Noise vector not the multiple Gauss (independent identical distributed complex gaussian) of independent same distribution, but, assuming that it is, the sub-optimality (sub-optimality) of the first detector block DET1 is in this approximate (approximation), and it allows strong complexity reduce and do not affect the order of diversity of described decoder output.

Assuming that Z ' ∈ Γ, wherein Z ' (i) is QAM modulation symbol, and Γ is the multiple QAM modulation space of s-dimension, and claims Γ T _ube satisfied point set, described " algebraic lattice simplification " technology is by acceptance point neighbouring drafting belongs to Ω _tusome Θ list and calculate the soft output of described first detector block DET1 by edge calculation on the list.Because equivalent channel is orthogonal, little point list just can reach the performance of best exhaustive marginalisation.Along with the spectrum efficiency of system increases, the minimizing of complexity too increases.

Therefore, subsequently, even if the channel formed by precoding beam at multiple spatial flow and when being launched, always can reaching the target diversity order of system, and be not to reach when this channel at the Wave beam forming of non-precoding.

Claims (23)

1. transmit data to the system of receiver from transmitter for the mimo channel that formed by launching beam, described transmitter comprises at least two transmitting antennas, follow the encoder (ENC) of error correcting code structure and implement the beam-forming device (TxBF) of launching beam formation technology, the decoder that described receiver comprises at least two reception antennas and limits according to described error correcting code structure, wherein said system comprises described transmitter and described receiver, it is characterized in that the code rate Rc of encoder (ENC), the selection of the parameter of mimo channel that described launching beam is formed and the target diversity order of system is complementary,

Wherein said transmitter (TRD) also comprises digital modulator (MOD), coded bit stream is converted to modulation symbol spatial flow by described digital modulator, the output of described digital modulator (MOD) is modulation symbol, and described modulation symbol defines N _sthe vector Z of individual modulation symbol Z (i), wherein i=1 ..., N _s, described transmitter (TRD) comprises algebraic linear precoder (ALP) further, and its parameter s is called as precoding size, and it meets s≤N _s, in order to by the N in same vector Z _ss modulation symbol linear combination in individual modulation symbol Z (i) together, is called hybrid modulation symbol, and in order to provide be associated with a transmission of the mimo channel formed by launching beam by N _sthe output vector X that individual pre-coded symbols X (i) forms,

The N of wherein said output vector X _sindividual pre-coded symbols X (i) is obtained by the vector Z of modulation symbol and the product of complex linear matrix S, and described complex linear matrix S is provided by following:

Wherein P ₁and P ₂it is permutation matrix, and S ' meets following s × s matrix, if the vector be namely made up of s modulation symbol and matrix S ' be multiplied, sent by the diagonal band noise Rayleigh fading channel being of a size of s × s, and decoded by maximum likelihood decoder, then the order of diversity of performance equals s.

2. system according to claim 1, permutation matrix P ₁be selected as equaling unit matrix, permutation matrix P ₂be chosen as the pre-coded symbols [X (1), the X (N that make output vector X _s-s+2) ..., X (N _s)] be the modulation symbol [Z (1) for combining of the described vector Z of modulation symbol, ..., Z (s)] linear combination, other pre-coded symbols of such output vector X is the modulation symbol that do not combine and meets [X (2), ..., X (N _s-s+1)]=[Z (s+1) ..., Z (N _s)].

3. according to the system in claim 1-2 described in any one, described receiver (RCV) comprises detector (DET), and the output of described detector is to coded-bit c _jestimation, it is characterized in that: by assemble described digital modulator (MOD), described launching beam formed mimo channel and described detector (DET) limit equivalent channel model (BIBOCH), the input of described equivalent channel model is coded-bit c _j, the output of described equivalent channel model is to coded-bit c _jsoft estimation or firmly estimate, described equivalent channel model is counted as at receiver place comprising nested block attenuation channel NBFCH, and its Mathematical Modeling is made up of block cascade and has five parameters, is respectively: the quantity N of spatial flow _s≤ min (N _t, N _r), the quantity N of transmitting antenna _t, reception antenna quantity N _r, the set D={N of order of diversity to be associated with the equivalent channel coefficient of each piece _tn _r, (N _t-1) (N _r-1) ..., (N _t-N _s+ s) (N _r-N _s+ s) } and the lengths sets of block the quantity of cascade block equals N _sand wherein L is the coded-bit c of each code word _jquantity, the length LB (i) of each piece is derived by the quantity of the bit of each modulation symbol be associated with the i-th spatial flow, the decay stochastic variable be associated with i-th piece is defined as the combination of subset ∑ (i) be made up of the individual independent random variable of D (i), wherein, integer value D (i)≤N _tn _r, make its implication is and supposition D (1)=N _tn _rthere is the highest order of diversity, the code rate R of described encoder (ENC) _c, described nested block attenuation channel NBFCH five Parameter N _s, N _t, N _r, D, LB, precoding size s and accessible aims of systems order of diversity δ are by following relational links:

δ (s)=D (i), wherein i by provide,

Wherein | D| is the radix of set D.

4. system according to claim 3, described precoding size equals 1, it is characterized in that: the code rate Rc of described encoder (ENC), five Parameter N of described nested block attenuation channel NBFCH _s, N _t, N _r, D, LB and accessible aims of systems order of diversity δ are by following relational links:

δ (s)=D (i), wherein i by provide

Wherein | D| is the radix of set D,

D={N _tn _r, (N _t-1) (N _r-1) ..., (N _t-N _s+ 1) (N _r-N _s+ 1) } and LB={LB (1) ..., LB (N _s).

5. system according to claim 3, be characterised in that: identical modulation is used for each spatial flow, all pieces that relate to the NBFCH of described equivalent channel have identical length LB (i), and system diversity degree δ (s) observed in the output of receiver is by providing as follows:

6. system according to claim 3, is characterised in that: described precoding size s is selected to obtain target diversity order δ according to following equation _t:

7. according to the system in claim 1-2 described in any one, described receiver (RCV) comprises detector (DET), and the output of described detector is to coded-bit c _jestimation, be characterised in that: described detector (DET) comprises the first detector block DET1 and the second detector block DET2, described first detector block is intended to the coded-bit c estimating to be associated with s the modulation symbol combined of the vector Z of modulation symbol Z (i) carried by the vectorial Y received _j, described second detector block is intended to estimate the N with the described vector Z of modulation symbol _sthe coded-bit c that-s the modulation symbol do not combined is associated _j.

8. system according to claim 7, is characterised in that: the output estimation about described first detector block DET1 obtains independent of the output estimation of described second detector block DET2.

9. system according to claim 8, is characterised in that: described first detector block DET1 provides soft output estimation, to the coded-bit c be associated with s composite symbol _jdescribed soft output estimation consider the coded-bit c with described s hybrid modulation symbol _jthat be associated and prior probability π (c that is that provided by the output of decoder (DEC) _j).

10. system according to claim 9, is characterised in that: described second detector block DET2 process N _sthe N of-s the modulation symbol do not combined _s-s independent detection.

11. systems according to claim 7, are characterised in that: described first detector block DET1 allows the order of diversity recovering to be brought by described algebraic linear precoder (ALP).

12. systems according to claim 11, are characterised in that: to the coded-bit c be associated with described s hybrid modulation symbol _jestimation preferably based on by the following imperfect likelihood probability p (Y provided _{s, 0}| Z _{s, 0}) function calculate:

Wherein N ₀noise variance, Z _{s, 0}be the vector Z of modulation symbol, the wherein said modulation symbol do not combined equals 0, i.e. Z _{s, 0}=[Z (1) ..., Z (s), 0 ..., 0], and Y _{s, 0}=[Y (1) ..., Y (s), 0 ..., 0] be the vector received, wherein last (N _r-s) individual value equals 0, and wherein V is unitary matrice, V ⁺be the transpose conjugate of matrix V, Δ is the diagonal angle rectangular matrix of diagonal values, and described diagonal values is the relevant non-identity distribution stochastic variable sorted with size descending.

13. systems according to claim 7, are characterised in that: described second detector block DET2 is to provide the linear detector of soft output estimation or hard output estimation.

14. systems according to claim 13, are characterised in that: described second detector block DET2 provides soft output estimation, to N _sthe coded-bit c that-s the modulation symbol do not combined is associated _jdescribed soft output estimation consider and described N _sthe coded-bit c of-s the modulation symbol do not combined _jthat be associated and prior probability π (c that is that provided by the output of decoder (DEC) _j).

15. systems according to claim 11, are characterised in that: the vectorial Y received at receiver place is provided by following:

Wherein Δ is the diagonal angle rectangle N of diagonal values _s× N _rmatrix, described diagonal values is the relevant non-identity distribution stochastic variable sorted with size descending, and wherein the matrix H of channel is written as H=U Δ _hv, wherein U is N _t× N _tunitary matrice, and V is N _r× N _runitary matrice, Δ _hthe diagonal angle rectangle N of diagonal values _t× N _rmatrix, described diagonal values is the relevant non-identity distribution stochastic variable sorted with size descending, and beam forming matrix T is by U ⁺front N _sthe N that individual row builds _s× N _tmatrix, V ⁺the transpose conjugate of matrix V, U ⁺the transpose conjugate of matrix U, vectorial N and N ₂be white Gauss noise vector, the vector of described reception is by given as follows:

YV ⁺＝[Z′S′M+N′；Z″M′+N″]

Wherein Z=[Z '; Z "] is the vector of modulation symbol, M and M ' is s × s diagonal angle rectangular matrix, and N ' and N, and " be all noise vector, the feature of system is that the first detector block DET1 calculates Z ' S ' M+N ' to the coded-bit c be associated with Z ' _jthe conversion of estimation, and described second detector block DET2 calculate Z " M '+N " to Z " the coded-bit c be associated _jthe conversion of estimation.

16. systems according to claim 15, are characterised in that: described s × s matrix M is decomposed by M=Ω Ψ Φ, and wherein Ω is s × s diagonal matrix, makes S ' Ω=T _us ', and T _ube s × s substrate variations matrix, this substrate variations entry of a matrix is complex integers, and diagonal angle s × s matrix Ψ has positive real diagonal element, and diagonal angle s × s matrix Φ has unimodular complex element, at receiver place, Ω is selected with optimization system performance, and the signal YV received ⁺by by the following matrix F filtering provided:

Wherein I is unit matrix, described first detector block DET1 compute vector to the coded-bit c be associated with Z ' _jhard estimation or the conversion of soft estimation, and described second detector block DET2 calculate Z " M '+N " to Z " the coded-bit c be associated _jhard estimation or the conversion of soft estimation.

17. according to the system in claim 1-2 described in any one, be characterised in that: described transmitter (TRD) comprises bit interleaver (INT), described receiver (RCV) comprises deinterlacer (DINT), according to error correcting code structure design bit interleaver (INT) to guarantee that the performance of the output at the decoder be associated (DEC) shows the target diversity order that can obtain.

18. transmit data to the device of receiver for the mimo channel formed by launching beam, described receiver comprises at least two reception antennas, described device (TRD) comprises at least two transmitting antennas, follows the encoder of error correcting code structure (ENC), implement the beam forming device (TxBF) of launching beam formation technology and coded bit stream is converted to the digital modulator (MOD) of spatial flow of modulation symbol, the output of described digital modulator is modulation symbol, and described modulation symbol defines N _sthe vector Z of individual modulation symbol Z (i), wherein i=1 ..., N _sdescribed receiver (RCV) comprises the decoder (DEC) limited according to error correcting code structure, be characterised in that described device (TRD) also comprises algebraic linear precoder (ALP), its parameter s is called as precoding size, and it meets s≤N _s, in order to by the N in same vector Z _ss modulation symbol linear combination in individual modulation symbol Z (i) together, is called hybrid modulation symbol, and in order to provide be associated with a transmission of the mimo channel formed by launching beam by N _sthe output vector X that individual pre-coded symbols X (i) forms,

The N of wherein said output vector X _sindividual pre-coded symbols X (i) is obtained by the vector Z of modulation symbol and the product of complex linear matrix S, and described complex linear matrix S is provided by following:

Wherein P ₁and P ₂it is permutation matrix, and S ' meets following s × s matrix, if the vector be namely made up of s modulation symbol and matrix S ' be multiplied, sent by the diagonal band noise Rayleigh fading channel being of a size of s × s, and decoded by maximum likelihood decoder, then the order of diversity of performance equals s.

19. devices according to claim 18, permutation matrix P ₁equal unit matrix, it is characterized in that permutation matrix P ₂be selected as the pre-coded symbols [X (1), the X (N that make output vector X _s-s+2) ..., X (N _s)] be the described vector Z of modulation symbol for combine modulation symbol [Z (1) ..., Z (s)] linear combination, other pre-coded symbols of such output vector X meets [X (2) ..., X (N _s-s+1)]=[Z (s+1) ..., Z (N _s)].

20. for the mimo channel that formed by launching beam from comprising N _tthe transmitter of>=2 transmitting antennas receives the device of data vector Y, the decoder (DEC) that described device (RCV) comprises Nr>=2 reception antenna, limit according to error correcting code structure, to export be the detector (DET) of estimation to coded-bit, s component in the vectorial Y of described reception be transmitting by N _sthe linear combination of s modulation symbol in the vector Z that individual modulation symbol Z (i) forms, wherein i=1 ..., N _s, s meets s≤N _sinteger value, it is characterized in that: described detector (DET) comprises the first detector block DET1 and the second detector block DET2, described first detector block is intended to estimate and the coded-bit that the modulation symbol of the vector Z of the modulation symbol carried by the described data vector Y received involved in s linear combination is associated, described second detector block is intended to the coded-bit estimating to be associated with the modulation symbol of the described vector Z of the modulation symbol do not related in described linear combination

Wherein the linear combination of modulation symbol is obtained by algebraic linear precoder (ALP) at transmitting pusher side,

Wherein the output of precoder is by N _sthe output vector X that individual pre-coded symbols X (i) forms, the N of described output vector X _sindividual pre-coded symbols X (i) is obtained by the vector Z of modulation symbol and the product of complex linear matrix S, and described complex linear matrix S is provided by following:

Wherein P ₁and P ₂it is permutation matrix, and S ' meets following s × s matrix, if the vector be namely made up of s modulation symbol and matrix S ' be multiplied, sent by the diagonal band noise Rayleigh fading channel being of a size of s × s, and decoded by maximum likelihood decoder, then the order of diversity of performance equals s.

21. devices according to claim 20, it is characterized in that: described first detector block DET1 allows the order of diversity recovering to be brought by described algebraic linear precoder, and described second detector block DET2 is to provide the linear detector of soft output estimation or hard output estimation.

22. devices according to claim 21, are characterised in that: the vectorial Y of reception is provided by following:

Wherein Δ is the diagonal angle rectangle N of diagonal values _s× N _rmatrix, described diagonal values is the relevant non-identity distribution stochastic variable sorted with size descending, and wherein the matrix H of channel is written as H=U Δ _hv, wherein U is N _t× N _tunitary matrice, and V is N _r× N _runitary matrice, Δ _hthe diagonal angle rectangle N of diagonal values _t× N _rmatrix, described diagonal values is the relevant non-identity distribution stochastic variable sorted with size descending, and beam forming matrix T is by U ⁺front N _sthe N that individual row builds _s× N _tmatrix, V ⁺the transpose conjugate of matrix V, U ⁺the transpose conjugate of matrix U, vectorial N and N ₂be white Gauss noise vector, the vector of described reception is by given as follows:

YV ⁺=[Z ' S ' M+N '; Z " M '+N "] wherein Z=[Z '; Z "] is the vector of modulation symbol; M and M ' is s × s diagonal angle rectangular matrix; and N ' and N " is noise vector, first detector block DET1 calculates the Z ' S ' M+N ' conversion to the estimation to the coded-bit be associated with Z ', and described second detector block DET2 calculating Z " M '+N " to the Z " conversion of the estimation of the coded-bit be associated.

23. devices according to claim 22, are characterised in that: described s × s matrix M is decomposed by M=Ω Ψ Φ, and wherein Ω is s × s diagonal matrix, makes S ' Ω=T _us ', and T _ube s × s substrate variations matrix, this substrate variations entry of a matrix is complex integers, and diagonal angle s × s matrix Ψ has positive real diagonal element, and diagonal angle s × s matrix Φ has unimodular complex element, at receiver place, Ω is selected with optimization system performance, and the signal YV received ⁺by by the following matrix F filtering provided:

Wherein I is unit matrix, and described first detector block DET1 is by vector convert the hard estimation to the coded-bit be associated with Z ' or soft estimation to, and described second detector block DET2 Z " M '+N " is converted to Z " the hard estimation of the coded-bit be associated or soft estimation.